Only send final PUT on Azure profiles
Fix core dump when reading Azure key from environment variable
Improve interface between mod_http_cache and storage providers
Remove debug
Remove out-of-date TODO
Add new configuration to example
Changes arising from Christopher Rienzo's comments
Add function to pointers to the profile to allow for better integration with the storage providers.
Add function to pointers to the profile to allow for better integration with the storage providers
Remove old header file inclusion
The member can hangup the call while we are playing the outbound announce audio to the agent, this
patch detects this and properly hangup the agent channel to avoid keeping it up with
just a silence call since there's no active bridge.
The agent-bridge-end event now has CC-Agent-Bridged: true if the bridge successfully happens,
variable cc_agent_bridged was added so we can track this in the agent and member CDRs.
If you want to play an audio to the agent to indicate the member hangup the call before the bridge
set the variable cc_bridge_failed_outbound_announce, for example:
<action application="set" data="cc_bridge_failed_outbound_announce=tone_stream://%(250,250,425);loops=3"/>
or
<action application="set" data="cc_bridge_failed_outbound_announce=your-caller-has-gone-away-audio"/>
prior to calling callcenter application.
These variables are documented at https://freeswitch.org/confluence/display/FREESWITCH/mod_callcenter
Increase value of samples to skip at the beginning of new frame.
This increases the robustness against false detections observed
on German voicemail from Switzerland (see wav audio attached
to case FS-9447).
Allows customizing the banner displayed when video is muted, via conference
param 'video-mute-banner' per conference, or channel var 'video_mute_banner'
per user.
Foreground/background colors, font face/size, and text can all be customized.
* commit '6c3710df4dc14751b60ba8f8ed4780e0f6288ec8':
FS-9264: Introduce two new api calls named detect_audio and detect_audio_silence. The existing wait_for_silence call never actually waits for silence until it first detects non-silence. There is also no way to set an independent timeout for detecting both the non-silence and then silence. This causes problems when wait_for_silence is called on an already quiet channel. Splitting the function up into two separate calls with separate timeouts offers more flexibility.
The following variables are added:
execute_on_spandsp_tone_detect_<TONE NAME> : executes APP when tone is detected. For example, if you have a tone
named "SIT" defined in spandsp.conf.xml, the variable would be "execute_on_spandsp_tone_detect_SIT"
api_on_spandsp_tone_detect_<TONE NAME> : similar to execute_on except that it executes an API.
spandsp_tone_detect_stop_on_tone : default is false. If true, detector will stop once the first matching tone is detected.
Christopher Rienzo